17 Inferior Vena Caval Filters
Matthew Matson
M. Matson, MD
Royal London Hospital, Whitechapel Road, London, E1 1BB, UK
17.1
Introduction
Pulmonary embolism (PE) is a major cause of mor- tality, accounting for 30,000–40,000 deaths per year in the UK, exceeded only by coronary artery disease and malignancy. The mainstay of treatment is anti- coagulation, initially with intravenous unfraction- ated heparin or subcutaneous low molecular weight heparin, followed by warfarin. This treatment is usually very successful in preventing further pul- monary emboli, and allowing endogenous throm- bolytic mechanisms to disperse existing thrombus.
In a minority of patients, further intervention is required.
17.2 History
Just under two decades after Virchow in 1846 hypoth- esised that pulmonary embolism was the result of migration of clot from the lower limb veins, Trous- seau, in 1865, suggested that embolization to the pulmonary arteries could be prevented by interrupt- ing the vessels providing the route. This led, in the 1930s, to the development of the technique of femoral vein ligation, with or without clot removal from the femoral vein. However, ligation of the femoral vein was associated with a high incidence of lower limb swelling, and obviously was ineffective in preventing more central, or contralateral deep venous throm- bus (DVT) from migrating to the lungs. For these reasons, in the 1940s, the focus switched to ligation of the inferior vena cava (IVC). At around the same time, heparin and warfarin became available, and were introduced into clinical practice in conjunction with these techniques. However, ligation of the IVC resulted in an immediate significant drop in cardiac output of around 50% and a significant incidence of bilateral lower limb swelling of around 15%. The mortality was up to 15%. In addition, further pulmo- nary emboli could occur from migration of thrombus above the level of ligation, a risk minimised by ligat- ing the IVC just below the level of the renal veins.
Embolisation could also occur through the collateral veins that invariably developed following IVC liga- tion and could approach the diameter of the original IVC. Alternatives to ligation of the IVC were therefore developed, including surgical plication of the IVC and surgical clips, allowing partial interruption of blood flow. Whilst demonstrating an advantage over liga- tion, these procedures still necessitated laparotomy.
The late 1960s saw the introduction of the first IVC filter, the Mobin Uddin umbrella, consisting of six struts placed with the apex inferiorly supporting a silastic membrane. The Mobin Uddin umbrella could be inserted via open femoral venotomy. However, the device suffered from problems with migration, caval occlusion and persisting pulmonary emboli.
CONTENTS
17.1 Introduction 145 17.2 History 145
17.3 Efficacy of IVC Filters 146
17.4 Indications and Contraindications 146 17.4.1 Accepted Indications 146
17.4.2 Uncertain Indications 147 17.4.3 Prophylactic Filters 147 17.4.4 Supra-renal Filters 147 17.5 Choice of Filter 147 17.5.1 The Ideal Filter 147
17.5.2 Currently Available Filters 147 17.5.2.1 Permanent Filters 147 17.5.2.2 Temporary Filters 148 17.5.2.3 Retrievable Filters 148 17.6 Insertion Technique 149 17.6.1 Pre-procedural Imaging 149 17.6.2 Filter Placement 149 17.6.3 Retrieval Technique 149 17.7 Complications 149 17.8 Conclusion 150 References 151
The next significant development was the release of the Greenfield filter. Initially also requiring sur- gical venotomy, since its release it has undergone modifications to allow percutaneous implantation to reduce complications of insertion site thrombo- sis. Other Changer have reduced the incidence of caval penetration by filter struts. The Greenfield filter remains available as a permanent device, with good follow-up data.
17.3
Efficacy of IVC Filters
The vast majority of the literature on IVC filters con- sists of observational studies, usually retrospective, of various devices. A number of recent, extensively referenced reviews have been published, critically evaluating the evidence for their use (Kinney 2003;
Hann and Streiff 2005).
The only prospective randomised controlled study to investigate the effectiveness of caval filters is the PREPIC study (Decousus et al. 1998). In this study, 400 patients with proven DVT seen on venog- raphy were assigned to receive either anticoagula- tion alone, in the form of unfractionated heparin or enoxaparin followed by oral anticoagulants for a minimum of 3 months, or anticoagulation and an IVC filter. At 10 days, the incidence of PE was sig- nificantly lower in the filter group, but there was no significant difference in the incidence of fatal PE.
No significant difference was seen in overall mortal- ity between the groups at 10 days or at follow-up of 2 years, but at 2 years there was a significant increase in the number of DVTs seen in the filter group.
The PREPIC study has been criticised; the indi- cations for filter placement were not the generally
accepted criteria and the number of patients recruited was half the originally intended number. Given that DVT is commoner than PE, it is possible that with a larger study population, the difference in recurrent PE at 2 years may have reached significance in the same way that the difference in recurrent DVT did.
17.4
Indications and Contraindications
Table 17.2. Indications for IVC fi lter placement
Accepted Indications
PE and contraindication to, or serious complication of, anticoagulation
Recurrent PE despite adequate anticoagulation Uncertain Indications
Free-fl oating thrombus
Before thrombolysis/mechanical thrombectomy Prophylactic fi lters
(trauma, high-risk orthopaedic surgery, bariatric surgery)
17.4.1
Accepted Indications
The firmest and commonest indication for IVC filter insertion is pulmonary embolism with a contrain- dication to anticoagulation or severe complication of anticoagulation. Examples would include pul- monary embolism and recent intracranial haemor- rhage or active gastrointestinal bleeding.
The other generally accepted indications for an IVC filter are pulmonary embolism despite adequate anticoagulation or propagation of DVT despite anti- coagulation. The diagnosis of pulmonary embolism while on treatment should be confirmed by investi- gation, since only a minority of symptomatic patients will have objective evidence of further emboli. If confirmed, before filter placement is contemplated other measures first require consideration, such as ensuring compliance with anticoagulation, increas- ing oral anticoagulation to a higher target therapeu- tic ratio or transferring to a different anticoagulant.
True failure of anticoagulation is rare.
Propagation of thrombus during anticoagulation occurs in 5%–10% of patients. However, the link between propagation of thrombus and embolisation to the lungs is not firmly established. Moreover, the placement of a filter does not prevent, and may exac- erbate thrombus propagation. Therefore, the initial
Table 17.1. Results of PREPIC study
At 12 days:
Signifi cant reduction in recurrent PE at 12 days in fi lter group
Non-signifi cant reduction in fatal recurrent PE at 12 days in fi lter group
No difference in mortality At 2 years:
Non-signifi cant reduction in recurrent PE in fi lter group Non-signifi cant reduction in fatal recurrent PE in fi lter group
Signifi cant increase in recurrent DVT in fi lter group No difference in mortality
treatment strategy in this scenario should be to opti- mise anticoagulant therapy.
17.4.2
Uncertain Indications
The remaining proposed indications for IVC place- ment are more contentious, often described as “rela- tive” or “possible” indications. They may perhaps be best regarded as indications for which there is no good evidence at present.
Some authors have advocated the presence of “free- floating” thrombus in the proximal lower limb veins or IVC as an indication for IVC filter, citing a very high incidence of pulmonary embolism of up to 60%.
However, in some of the studies the diagnosis of free- floating thrombus was made after PE had occurred and another prospective study (Pacouret et al. 1997) showed no increased risk of PE in this patient group.
Similarly there are advocates of filter placement to prevent PE when performing thrombolysis or mechanical thrombectomy of proximal venous thrombus, while others consider it safe to perform these procedures without protection.
17.4.3
Prophylactic Filters
Use of IVC filters has been recommended in certain high-risk groups of patients, including trauma patients, patients undergoing orthopaedic procedures and those having bariatric surgery for morbid obesity.
In patients with polytrauma, there is a high inci- dence of venous thrombo-embolic disease, with observed rates of deep venous thrombosis of up to 58% and pulmonary embolism in up to 4%. There is some evidence that heparin prophylaxis is not completely effective in this group and studies com- paring filter use with historical controls or low dose heparin suggest an advantage to filter use. However, the level of evidence is not sufficient to recommend their routine use. Instead, they should be considered in very high risk groups, such as complex orthopae- dic lower limb injury and spinal injury.
In orthopaedic procedures, there is an incidence of venous thromboembolism of up to 20% without prophylaxis, particularly following total hip or knee replacement. However, that risk can be substantially lowered by the correct prescribing of low molecular weight heparin, such that routine IVC filters are not indicated prophylactically.
An incidence of PE of up to 3% is seen in patients undergoing bariatric surgery, but with a high associ- ated mortality rate, most likely because the underly- ing morbid obesity results in reduced cardiopulmo- nary reserve. This has led to some surgeons using prophylactic IVC filters, but there is no evidence for their use in this group and, as for high-risk ortho- paedic procedures, heparin prophylaxis probably remains a more logical choice.
17.4.4
Supra-renal Filters
The commonest indication for supra-renal place- ment of a filter is thrombus extending up into the IVC, such that there is no room for infra-renal place- ment. If a filter is required in a pregnant woman, then the supra-renal location should be used. Some workers have extended this argument to advocate supra-renal placement in all women with the capac- ity for future pregnancy, though this is less impor- tant if a non-permanent filter is used.
17.5
Choice of Filter
17.5.1
The Ideal Filter
The concept of the “ideal filter” is useful as a com- parison for the currently available devices. The ideal filter is cheap and easy to insert and reposition through a small introducer system, so as never to cause entry point thrombosis. It has a 100% capture rate of emboli without causing impedance to flow and is atraumatic to the IVC with a 0% incidence of caval occlusion. It does not migrate and may be retrieved easily following an indefinite period. It is biocompatible and MRI compatible.
17.5.2
Currently Available Filters
17.5.2.1
Permanent Filters
There are currently five permanent filter devices
available in the UK: the Greenfield filter (Boston
Scientific, Natick, Mass, USA)), the Bird’s Nest filter
(Cook, Bloomington, Ind, USA), the LGM Vena-Tech filter (Braun, E-Z-EM, Boulogne, France), the Simon Nitinol filter (Bard, Temper, Arizona, USA) and the TrapEase filter (Cordis, Johnson + Johnson, New Brunswick, New Jersen, USA). These devices vary in their appearance, deployment mechanisms and construction materials but there are no controlled trials to show that any one device is superior to the others. The greatest volume of data exists for the Greenfield filter. The Bird’s Nest filter is perhaps more complex to deploy than the others, but is the filter of choice for patients with a large diameter caval lumen, up to 40 mm.
17.5.2.2
Temporary Filters
As stated above, the major indication for IVC filter insertion is contraindication to anticoagulation. In many instances, the contraindication to anticoagu- lation may be temporary: for example, the require- ment for surgery. From the PREPIC study there is evidence of an increased incidence of DVT in patients with a filter in situ. Temporary filters were developed as an attractive means of avoiding the long-term complications of filters. The disadvantage of these purely temporary filters is that the introducing cath- eter remains in place, which risks the catheter and filter becoming displaced, restricts movement of the patient and carries a risk of infection.
Recent years have, for these reasons, seen the withdrawal of most purely temporary filters in favour of retrievable filters.
17.5.2.3
Retrievable Filters
Retrievable filters offer the advantages of tempo- rary filters, without the disadvantages mentioned above. They may also be left in place to act as per- manent devices. The potential disadvantage is that they require two separate procedures, one to insert the device and one to remove it. Currently available retrievable filters in the UK are the Gunther Tulip filter (Cook), the Optease filter (Cordis), the ALN filter (Pyramed) and the Recovery filter (Bard).
The Gunther Tulip vena cava filter consists of a stainless steel, half basket with four wires meeting at the apex and smaller wires attached to the main wires giving a ‘tulip’ appearance. It can be retrieved via an 11-F coaxial sheath by snaring a hook attached
at the apex of the device. It is recommended by the manufacturer that it is retrieved within 10 days, after which time it becomes a permanent filter. In practice, a number of workers have reported its safe retrieval after longer periods of time.
The Optease filter is a variant of the permanent Trapease filter, but has a hook for snare retrieval. It may be delivered via either femoral, jugular or ante- cubital vein through a 6-F system. It is MRI compat- ible. It can be used in IVCs up to 30 mm in diameter.
It may be left in situ as a permanent device, but if it is to be retrieved, the maximum recommended dura- tion is 12 days.
Filters that can be left in situ for 10–12 days are ideal for situations where the contraindication to anticoagulation is short term, such as uncompli- cated surgery. There are, however, a number of situa- tions where the contraindication to anticoagulation is likely to be more prolonged: for example, a patient with major trauma or following neurosurgery may well not be recovered and ambulant within the 12- day period. For these indications, the concept of a filter that may be left in situ for longer, indefinite periods before removal is attractive. Two filters, the Recovery and the ALN filter, are licensed for long- term implantation prior to removal.
The ALN filter is a stainless steel alloy device with six short legs to fix it to the IVC wall and three long legs to ensure a central position within the IVC. It can be deployed from a jugular, antecubital or femo- ral venous access through a 7-F delivery sheath and retrieved from the jugular route through a 9-F sheath.
Pieri et al. (2003), Barral et al. (2003) and Pancione and Mecozzi (2004) have presented abstracts report- ing 100% retrieval rates in a total of 48 patients after a period of 11–192 days. Imberti et al. (2005) prospec- tively studied 30 patients who received ALN filters, in whom 18 retrievals were attempted. Successful retrieval was achieved in 78% of cases after a median interval of 123 days (range 30–345). Interestingly, the success rate was 100% in filters that had been in situ for less than 3 months, but was only 50% in the remaining eight patients, where adherence to the IVC wall prevented retrieval in three patients, and tilting prevented retrieval in one patient.
The Recovery filter is a nitinol device with six short
legs and six hooked long legs. It is delivered through
a 6-F system only from the femoral route, and may
be retrieved from the jugular route, using a retrieval
cone with nine metal claws delivered through a 10-
F system. It is MRI compatible. Asch et al. (2002)
studied its use in 32 patients, attempting retrieval in
24 patients. Retrieval was successful in 100% of cases
after a mean period of 53 days (range 5–134), even when there was capture of large thrombus.
17.6
Insertion Technique
17.6.1
Pre-procedural Imaging
The IVC is formed by the union of the common iliac veins at the L4/5 level. It ascends to the right of the aorta, receiving tributaries from lumbar veins, renal veins and hepatic veins before draining into the right atrium.
The most usual site for placement of an IVC filter is infra-renally. Many devices are recommended to be placed ideally with the cranial tip at the level of the renal veins, where there is increased flow, to minimise the risk of caval occlusion.
Placement is normally monitored using venog- raphy and fluoroscopy, although grey-scale ultra- sound, duplex scanning and intravascular ultra- sound have all been reported. The pre-procedural evaluation of anatomy has to assess the diameter of the IVC and accurately localise the number and levels of the renal veins. Anomalies of the venous anatomy are relatively common. Duplication of the IVC (0.2%–3%) may require two filter devices to
be used, or at least the selection of the correct IVC (Fig. 17.1). A retro-aortic left renal vein (1.8%–2.4%) usually enters more inferiorly and requires lower placement of the filter. Multiple renal veins may communicate at the renal hilum, offering a bypass route for emboli; this is particularly true in the case of circumaortic left renal vein (1.5%–8.7%).
In addition, imaging also serves to assess the extent of thrombus in the venous system and allow for planning of site of delivery access and desired position of the filter. For example, thrombus extend- ing into the iliac veins would be an indication for trans-jugular placement, and thrombus extending into the IVC up to the renal veins is an indication for supra-renal placement of the filter.
17.6.2
Filter Placement
The deployment of the filter is usually straightfor- ward. The manufacturers’ instructions are always comprehensive but can be, in the author’s opinion, somewhat technical and confusing; when using a device for the first time it is useful to have an expe- rienced person present to help and advise.
17.6.3
Retrieval Technique
Before removal of a retrievable filter a cavagram is required to assess filter position and exclude signifi- cant thrombus within it. If thrombus is present, a further period of anticoagulation may be required, if time allows, or thrombolysis may be considered, accepting the inherent risks and assuming no con- traindication. Retrieval into a large sheath is appro- priate for small trapped clots, again with consid- eration of the risks for the patient of inadvertent emboli. All retrievable filters may be left in situ as permanent devices.
The retrieval technique itself is essentially that of foreign body retrieval and should be performed according to the manufacturers’ instructions.
17.7
Complications
Complications of IVC filter placement can be divided into immediate and longer term. Immediate include
Fig. 17.1. Cavagram demonstrating duplicated IVC
access site complications associated with central venous puncture, such as inadvertent arterial punc- ture or pneumothorax. With early devices, access site thrombosis was a common problem, but this appears to be less so with modern devices. It has even been suggested that the rate can be lowered by less enthusiastic manual compression following the procedure. A number of operator error complica- tions have been described, including incorrect sizing of the IVC for the device, deployment of a femo- ral device via the jugular route and deployment of devices into renal veins, gonadal veins, iliac veins and the aorta (Kaufman et al. 1995). These com- plications can be easily avoided by careful choice of device and appropriate imaging.
Filter migration, defined as 2 cm movement of the device, was similarly seen with early devices. It is now rare (<1%) with modern filters, mostly as a result of hooks being incorporated into them. The downside of this is an increased incidence of caval penetration by the hooks. This is seen not uncom- monly in patients followed up on CT or venography, though clinically significant perforation into the aorta, duodenum or other retroperitoneal struc- tures is fortunately rare.
Filter fracture has led to withdrawal of some devices, such as the Antheor filter (Fig. 17.2), but the fracture rate with other devices, such as the Green- field filter, appears to be very low. The incidence of long term fractures with the newer devices has yet to be established, but the incidence in the short and medium term is low.
IVC thrombosis is an important long term com- plication of IVC filters. It is generally reported with an incidence of around 2%–10%, although one study found an incidence as high as 33% at 9-month follow- up. About half of patients who have IVC thrombosis develop symptoms, such as lower limb swelling. IVC thrombosis may involve the renal veins, though this in itself may not alter renal function.
From the PREPIC study, we know that DVT occurs more frequently in patients with IVC filters. Long- term sequelae of DVTs include post-phlebitic limb, with deep venous incompetence, limb swelling and venous ulceration.
17.8 Conclusion
There is evidence that IVC filters reduce the risk of pulmonary embolism in the short term in patients
Fig. 17.2. Fracture of an Antheor fi lter
Fig. 17.3. a Gunther Tulip fi lter in thrombosed IVC. b More cra- nial section demonstrating associated renal vein thrombosis
a
b
with proximal deep venous thrombosis. They should not be considered a substitute for anticoagulation:
unlike anticoagulants, they do not prevent propaga- tion of DVTs, they do not help prevent long-term sequelae of DVT such as post-phlebitic limb and they increase rather than decrease the risk of recur- rent DVT. The best accepted indication for their use is pulmonary embolism with a contraindication to anticoagulation. These contraindications are often temporary, such as bleeding peptic ulcer or surgery.
The development of retrievable filters that may be left in situ for a number of weeks before removal may avoid their long-term complications, such as recurrent DVT and caval occlusion. The continued improvements in filter design will no doubt be par- alleled by advances in anticoagulant medication, such that the place for IVC filters will have to be continually re-evaluated.
References
Asch M (2002) Initial experience in humans with a new retriev- able inferior vena cava filter. Radiology 225:835–844 Barral F, Tardy B, Guillot K et al (2003) Clinical experience with
optional cava filters. J Thromb Haemost 1:OC441
Decousus H, Leizorovicz A, Parent F et al for The Prévention du Risque d’Embolie Pulmonaire par Interruption Cave Study Group (1998) A clinical trial of vena caval filters in the pre- vention of pulmonary embolism in patients with proximal deep-vein thrombosis. N Engl J Med 338:409–415
Hann CL, Streiff MB (2005) The role of vena cava filters in the management of venous thromboembolism. Blood Rev 19:179–202
Imberti D, Bianchi M, Farina A et al (2005) Clinical experi- ence with retrievable vena caval filters: results of a prospec- tive observational multicenter study. J Thromb Haemost 3:1370–1375
Kaufman JA, Geller SC, Rivitz SM et al (1995) Operator errors during percutaneous placement of vena cava filters. AJR Am J Roentgenol 165:1281–1287
Kinney TB (2003) Update on inferior vena cava filters. J Vasc Intervent Radiol 14:425–440
Pacouret G, Alison D, Pottier JM et al (1997) Free-floating thrombus and embolic risk in patients with angiographi- cally confirmed proximal deep venous thrombosis. A pro- spective study. Arch Intern Med 157:305–308
Pancione L, Mecozzi B (2004) Permanent/removable vena cava filter ALN (France): our experience with 96 patients. Pro- ceedings of the 90th annual meeting of the Radiological Society of North America. Abstract SSJO3
Pieri S, Agresti P, Morruci M et al (2003) Optional vena cava filters: preliminary experience with a new vena cava filter.
Radiol Med 105:56–62
Terhaar OA, Lyon SM, Given MF et al (2004) Extended interval for retrieval of Gunther Tulip filters. J Vasc Intervent Radiol 15:1257–1262
